It is very difficult to trace the provenance of the data in PeriodicTable.jl to the primary source. This ambiguity hinders its use for serious scientific work. PT.jl should be able to reference the primary source of all data (publication reference preferably) . It is an important part of the reproducibility and traceability of scientific results. When the data changes, the semantic version number should be updated to reflect this is, from the perspective of a scientific user, a breaking change that I'd need to evaluate to decide whether I want to accept this change.
Currently, PeriodicTable.jl "sort of" addresses this question by referencing Wikipedia. However, this only passes the problem along. For one, since websites change, you need to reference the date the data was downloaded. Second, it makes tracing the provenance of a collection of values tedious and makes updating them difficult. PT.jl shouldn't be using Wikipedia. It should go directly to reference-able primary sources.
It is very difficult to trace the provenance of the data in PeriodicTable.jl to the primary source. This ambiguity hinders its use for serious scientific work. PT.jl should be able to reference the primary source of all data (publication reference preferably) . It is an important part of the reproducibility and traceability of scientific results. When the data changes, the semantic version number should be updated to reflect this is, from the perspective of a scientific user, a breaking change that I'd need to evaluate to decide whether I want to accept this change.
Currently, PeriodicTable.jl "sort of" addresses this question by referencing Wikipedia. However, this only passes the problem along. For one, since websites change, you need to reference the date the data was downloaded. Second, it makes tracing the provenance of a collection of values tedious and makes updating them difficult. PT.jl shouldn't be using Wikipedia. It should go directly to reference-able primary sources.